Novel high-efficiency 2D position-sensitive ZnS:Ag/6LiF scintillator detector for neutron diffraction

Abstract

Scintillator-based ZnS:Ag/⁶LiF neutron detectors have been under development at ISIS for more than three decades. Continuous research and development aim to improve detector capabilities, achieve better performance and meet the increasingly demanding requirements set by neutron instruments. As part of this program, a high-efficiency 2D position-sensitive scintillator detector with wavelength-shifting fibres has been developed for neutron-diffraction applications. The detector consists of a double scintillator-fibre layer to improve detection efficiency. Each layer is made up of two orthogonal fibre planes placed between two ZnS:Ag/⁶LiF scintillator screens. Thin reflective foils are attached to the front and back scintillators of each layer to minimize light cross-talk between layers. The detector has an active area of 192 × 192 mm with a square pixel size of 3 × 3 mm. As part of the development process of the double-layer detector, a single-layer detector was built, together with a prototype detector in which the two layers of the detector could be read out separately. Efficiency calculations and measurements of all three detectors are discussed. The novel double-layer detector has been installed and tested on the SXD diffractometer at ISIS. The detector performance is compared with the current scintillator detectors employed on SXD by studying reference crystal samples. More than a factor of 3 improvement in efficiency is achieved with the double-layer wavelength-shifting-fibre detector. Software routines for further optimizations in spatial resolution and uniformity of response have been implemented and tested for 2D detectors. The methods and results are discussed in this manuscript.

Keywords: SXD diffractometer; high efficiency; neutron detectors; neutron diffraction; scintillator detectors; wavelength-shifting fibre detector.

Figure 1

Sketch of the double-scintillator-fibre layer arrangement. A single layer consists of two orthogonal fibre planes (1 mm ø, 3 mm pitch) sandwiched between scintillator sheets (450 µm-thick) with reflective foils (50 µm-thick) and a 5 mm-thick B₄C plate to reduce neutron background.

Figure 2

Efficiency of different scintillator configurations. The current SXD detector efficiency is the theoretical calculation of absorption efficiency for the single scintillator sheet (black curve). The measured efficiency of the single-layer (0.45 mm-thick scintillator front and back) and the double-layer detectors are shown as the blue and red curves respectively.

Figure 3

Incident spectrum with respect to neutron wavelength of the double-layer prototype. The contribution of the two layers summed together is represented by the black curve. The red and blue distributions correspond to the front and back layers, respectively. The ratio between the distributions of the layers summed together (black curve) and the front layer (red curve) is depicted in the inset plot together with the exponential fit.

Figure 4

Laue diffraction image for an NaCl spherical crystal integrated in ToF with (a) the double-layer WLSF detector and (b) the current SXD detector. In order to obtain the same reflections on both detectors, the sample was measured at a fixed position and rotated 37.5° anti-clockwise for the current SXD detector measurements. The two plots are depicted in a sketch to illustrate the geometry and positions of the two detectors on the SXD instrument. The angle α is defined in equation (1). The colour bars represent the ToF integrated counts on a linear scale.

Figure 5

2D images of the Bragg reflections in (λ, θ) space for (a) the double-layer WLSF detector and (b) the current SXD detector. The colour bars represent counts on a linear scale.

Figure 6

Intensities of 2 2 and 4 4 Bragg reflections integrated over a 3 ×  5 detector pixel area around the (32, 28) Bragg spot recorded on the double-layer WLSF detector (red curve) and the corresponding peak recorded with the current SXD detector centred at pixel (33, 27) (blue curve). The intensities are normalized by measurement time and solid-angle coverage.

Figure 7

(a) 2D images of three 2 mm holes in a B₄C mask for the raw data (left) and the data analysed using the CoG position-interpolation algorithm (right). A sketch of the mask shows the positions of the three holes and highlights the effectiveness of the CoG algorithm. The colour bars represent the counts integrated by time on a linear scale. (b) Vertical profile of the three-hole mask, for the raw data (black) and for the CoG processed data (red). (c) Horizontal profile of the three-hole mask for the raw data (black) and for the CoG processed data (red). The three peaks corresponding to the three holes are clearly distinguished when the CoG is applied, but only two peaks are visible in the raw data.

Figure 8

Average counts for the X and Y fibre planes of the double-layer WLSF ZnS:Ag/⁶LiF detector as a function of lower-level discrimination (LLD) are shown as red and blue curves, respectively. The error bar represents the calculated standard deviation per LLD as ±σ.

Figure 9

Distribution of counts recorded with a V/Nb sample on the X and Y fibre planes for fixed LLD (black), variable LLD in (a) red and (d) blue, respectively, and the calculated event distribution fit in grey. Detector-uniformity variation for each fibre shown as the percentage deviation of the fixed LLD event distribution from the fit in black diamonds and of the variable LLD event distribution from the fit in (b) red and (e) blue circles for the X and Y fibre planes, respectively. The uniformity is improved by a factor of 5 when applying the variable LLD correction. (c) and (f) LLD offset value for each fibre from the fixed LLD = 200, corresponding to offset 0.